An exposure apparatus projects and exposes an image of a pattern of a mask (or a reticle) on a photosensitive substrate (e.g., wafer or a glass plate coated by a photoresist). More specifically, the exposure apparatus uses a projection optical system to project and expose the image in a photolithography process, in which semiconductor devices or the like are manufactured. The projection optical system is required to have a higher resolution due to the increasing level of integration of semiconductor devices or the like manufactured with the exposure apparatus.
To improve the resolution of the projection optical system, the projection optical system needs to shorten the wavelength λ of its illumination light (exposure light) and increase the numerical aperture NA at its image side. The resolution of the projection optical system is written as k*λ/NA (where k is a process coefficient). The image-side numerical aperture NA is written as n*sin θ, where n is the refractive index of a medium (usually gas, such as air) that arranged between the projection optical system and the photosensitive substrate, and θ is the maximum incident angle at which light enters the photosensitive substrate.
When the maximum angle θ is set larger in an effort to increase the image-side numerical aperture, the angle at which light is emitted from the projection optical system and the angle at which the light enters the photosensitive substrate increase. This increases reflection loss occurring at an optical surface. As a result, the projection optical system cannot obtain a large effective numerical aperture at its image side. One conventional technique for increasing the image-side numerical aperture is an immersion technique (refer to, for example, Patent Document 1), with which an optical path formed between a projection optical system and a photosensitive substrate is filled with a medium such as a liquid having a high refractive index.    Patent Document 1: International Patent Publication No. WO2004/019128